OverSeer: Multi-Agent World Models Beyond What Agents See

Hugging Face Daily Papers Papers
world-models multi-agent video-generation delta-vae representation-learning minecraft embodied-ai

Summary

OverSeer introduces a multi-agent world model that augments egocentric agent views with fixed auxiliary camera views acting as a world-centric representation, compressed via delta-VAE to cut token computation, showing improved consistency in Minecraft and Doom.

Despite the success of video world models, extending them to multiple agents is difficult because it requires generating several agent-specific video streams that are consistent in time and between agents. To address this gap, we introduce OverSeer, a multi-agent world model that augments egocentric agent views with auxiliary camera views. We suggest that, by fixing the auxiliary views in space, they act as a proxy for a world-centric representation: Unlike the agents' views, which can change entirely due to a simple change in agent position, changes in the auxiliary views directly map to localized changes in the world state. We show that these fixed views substantially help the model to maintain consistency. We also show that, due to their redundant and semi-static nature, the auxiliary views can be efficiently compressed with a delta-VAE that encodes only the changes relative to a static reference image, reducing the number of tokens required to process them. Experiments in Minecraft and Doom demonstrate improved world and agent consistency, with qualitative results on real-world multi-camera captures indicating that the approach is not tied to game engines. Our compression analysis further shows that delta-VAE substantially reduces the computational overhead of auxiliary-view generation. We conclude that these fixed views provide a simple and efficient way to improve multi-agent world models.
Original Article
View Cached Full Text

Cached at: 10/21/26, 02:22 PM

Paper page - OverSeer: Multi-Agent World Models Beyond What Agents See

Source: https://huggingface.co/papers/2610.28554

Abstract

OverSeer improves multi-agent world models by adding fixed auxiliary camera views that stabilize consistency and are compressed via delta-VAE to reduce computation.

Despite the success of video world models, extending them to multiple agents is difficult because it requires generating several agent-specific video streams that are consistent in time and between agents. To address this gap, we introduce OverSeer, amulti-agent world modelthat augments egocentric agent views withauxiliary camera views. We suggest that, by fixing the auxiliary views in space, they act as a proxy for aworld-centric representation: Unlike the agents’ views, which can change entirely due to a simple change in agent position, changes in the auxiliary views directly map to localized changes in the world state. We show that these fixed views substantially help the model to maintain consistency. We also show that, due to their redundant and semi-static nature, the auxiliary views can be efficiently compressed with adelta-VAEthat encodes only the changes relative to a static reference image, reducing the number of tokens required to process them. Experiments in Minecraft and Doom demonstrate improved world and agent consistency, with qualitative results on real-world multi-camera captures indicating that the approach is not tied to game engines. Our compression analysis further shows thatdelta-VAEsubstantially reduces the computational overhead of auxiliary-view generation. We conclude that these fixed views provide a simple and efficient way to improvemulti-agent world models.

View arXiv pageView PDFProject pageAdd to collection

Get this paper in your agent:

hf papers read 2610\.28554

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2610.28554 in a model README.md to link it from this page.

Datasets citing this paper2

#### liguang0115/overseer-minecraft Viewer• Updated17 minutes ago • 38.8k • 338 #### liguang0115/overseer-doom Viewer• Updated17 minutes ago • 19.8k • 246

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2610.28554 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

MultiWorld: Scalable Multi-Agent Multi-View Video World Models

Hugging Face Daily Papers

MultiWorld is a unified framework for multi-agent multi-view video world modeling that achieves accurate control of multiple agents while maintaining multi-view consistency through a Multi-Agent Condition Module and Global State Encoder.

World Observer: Joint Actor-Observer Generation for Persistent World Modeling

Hugging Face Daily Papers

World Observer introduces a joint actor-observer generation framework for video world models, using panoramic observers to keep tracking out-of-view objects so their state stays consistent when they re-enter the actor's view. The paper also proposes world-space metrics and a real/synthetic benchmark for evaluating out-of-view evolution.

Agora-1: The Multi-Agent World Model

Hacker News Top

Odyssey introduces Agora-1, a multi-agent world model that enables real-time shared simulations for multiple participants, demonstrated with a GoldenEye deathmatch game.

MASS: Multiplayer World Models with Authoritative Shared State

Hugging Face Daily Papers

This paper introduces MASS, a method for multiplayer world models that disentangles world dynamics from view rendering using an authoritative shared state, enabling scalable and consistent multi-agent simulation with up to 1,024 concurrent players.